ENV CHECK (embeddings + multiprocessing)¶

In [1]:
# figure settings
import scanpy as sc
import anndata as ad
import os
sc.set_figure_params(scanpy=True, dpi=80, dpi_save=250,
                     frameon=True, vector_friendly=True,
                     color_map="YlGnBu", format='pdf', transparent=False,
                     ipython_format='png2x')

DATADIR = "/storage/groups/ce01/workspace/mobisc_anna/processed_datasets/share_seq_brain"
#%ls $DATADIR
files = [os.path.join(DATADIR, 'old_anndata_brain_share_seq_geneactivity_rna_merged.h5ad')]

int_files = ['liger_jnmf_integration.h5ad', 
             'scJoint_integration.h5ad', 
             'seurat_cca_integration.h5ad', 
             'seurat_rpca_integration.h5ad']

DATADIR = "/storage/groups/ce01/workspace/mobisc_anna/integrated_outputs/share_seq_brain_shared_genes_full"
#%ls $DATADIR
for i in int_files:
    files.append(os.path.join(DATADIR, i))
    
data = ad.read(files[0])
sc.pp.pca(data)
sc.pp.neighbors(data)
sc.tl.umap(data)
sc.pl.umap(data, color='broad_annotations')

from multiprocessing import Process, Queue from time import sleep

def my_func(n, l): if n % 2 == 0: sleep(5) for i in l: print(f'Process {n}: {i}\n')

processes = [] my_list = range(1, 5) for i in range(8): p = Process(target=my_func, args=(i, my_list)) p.start() processes.append(p)

for p in processes: p.join()

FURTHER ANALYSIS¶

In [2]:
# Load packages 
import scanpy as sc
import anndata as ad
import os
import pandas as pd
import matplotlib.pyplot as plt
import scmoib
import numpy as np
import tqdm

During startup - Warning messages:
1: Setting LC_TIME failed, using "C" 
2: Setting LC_MONETARY failed, using "C" 
3: Setting LC_PAPER failed, using "C" 
4: Setting LC_MEASUREMENT failed, using "C"
In [3]:
# figure settings
sc.set_figure_params(scanpy=True, dpi=80, dpi_save=250,
                     frameon=True, vector_friendly=True,
                     color_map="YlGnBu", format='pdf', transparent=False,
                     ipython_format='png2x')

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scanpy/_settings.py:447: DeprecationWarning:

`set_matplotlib_formats` is deprecated since IPython 7.23, directly use `matplotlib_inline.backend_inline.set_matplotlib_formats()`

Load the processed data¶

DATADIR = "/storage/groups/ce01/workspace/mobisc_anna/processed_datasets/share_seq_brain"

%ls $DATADIR¶

files = [os.path.join(DATADIR, 'old_anndata_brain_share_seq_geneactivity_rna_merged.h5ad')]

int_files = ['liger_jnmf_integration.h5ad', 'scJoint_integration.h5ad', 'seurat_cca_integration.h5ad', 'seurat_rpca_integration.h5ad']

DATADIR = "/storage/groups/ce01/workspace/mobisc_anna/integrated_outputs/share_seq_brain_shared_genes_full"

%ls $DATADIR¶

for i in int_files: files.append(os.path.join(DATADIR, i))

ids = ['unintegrated', 'liger', 'scjoint', 'seurat_cca', 'seurat_rpca'] modes = ['raw', 'comps', 'comps', 'count', 'count'] col = ['broad_annotations'] * 5 reps = [None, 'X_inmf_embedding', 'X_scjoint_embedding', None, None] my_obj = scmoib.tl.MetricsCalculator() adata_list = []

In [4]:
DATADIR = "/mnt/znas/icb_zstore01/groups/ce01/workspace/mobisc_anna/processed_datasets/PBMC_10X/"
#%ls $DATADIR
files = [os.path.join(DATADIR, 'old_anndata_pbmc_10x_processed_merged_rna_first_full_feature.h5ad')]

int_files = ['liger/PBMC_10X_jNMF.h5ad', 
             'seurat_cca/pbmc_10x_seurat_cca_rna_first.h5ad', 
             'bindsc/pbmc_10x_bindsc_rna_first_full_feature.h5ad', 
             'scalex/RNA-ATAC_pbmc_10x_scalex_processed_full_feature.h5ad', 
             'seurat_rpca/pbmc_10x_seurat_rpca_rna_first.h5ad']

DATADIR = "/storage/groups/ce01/workspace/mobisc_anna/integrated_outputs/"
%ls $DATADIR
for i in int_files:
    files.append(os.path.join(DATADIR, i))
    
ids = ['unintegrated', 'liger', 'seurat_cca', 'bindsc', 'scalex', 'seurat_rpca']
modes = ['raw', 'comps', 'count', 'count', 'graph', 'count']
col = ['broad cell types', 'broad.cell.types'] + ['broad cell types'] * 4
reps = [None, 'X_iNMF', None, None, None, None]
my_obj = scmoib.tl.MetricsCalculator()
adata_list = []

PBMC_10x/  pbmc_jnmf.h5ad    scalex/       share_seq_brain_shared_genes_full/
bindsc/    pbmc_uinmf.h5ad*  seurat_cca/
liger/     scJoint/          seurat_rpca/
In [5]:
%%time


for i in tqdm.tqdm(files):
    pbmc = ad.read(i)
    adata_id = ids[files.index(i)]
    clust_id = 'louvain'
    ct_id = col[files.index(i)]
    bc_list = list(pbmc.obs.index)
    bc_list1 = bc_list[:len(bc_list) // 2]
    bc_list2 = bc_list[len(bc_list) // 2:]
    mode = modes[files.index(i)]
    rep = reps[files.index(i)]
    my_obj.check_anndata(pbmc, ct_id, mode=mode, use_rep=rep)
    print('-' * 10)
    my_obj.node_metrics(pbmc, adata_id, bc_list1, bc_list2, ct_id, n_jobs=8)
    my_obj.spec_dist(pbmc, adata_id=adata_id, n_metr=10)
    try:
        my_obj._accuracy_paired_omics(pbmc, adata_id, bc_list1, bc_list2, 'omic_batch', clust_id)
        my_obj._accuracy_paired_omics_per_cell_type(pbmc, adata_id, bc_list1, bc_list2, 
                                                    'omic_batch', clust_id, ct_id)
        my_obj.ami(pbmc, adata_id, ct_id, clust_id)
        my_obj.ari(pbmc, adata_id, ct_id, clust_id)
        my_obj.nmi(pbmc, adata_id, ct_id, clust_id)
        my_obj.homogeneity(pbmc, adata_id, ct_id, clust_id)
    
    except ValueError as err:
        print(f"Trouble with {adata_id}")
        print(err.args[0])

    my_obj._graph_connectivity(pbmc, adata_id, ct_id)
    try:
        my_obj.silhouette(pbmc, adata_id, batch_key='omic_batch', embed='X_umap')
        my_obj.silhouette_batch(pbmc, adata_id, batch_key='omic_batch', cell_label=ct_id, embed='X_umap')
    except:
        print('Silhouette failed')
    my_obj._average_dist_between_matching_barcodes(pbmc, adata_id, bc_list1, bc_list2, metric='cosine')
    my_obj.isolated_labels_score(pbmc, adata_id, ct_id, 'omic_batch', 'X_umap')
    adata_list.append(pbmc)
my_obj.get_df()

  0%|                                                     | 0/6 [00:00<?, ?it/s]
Running pre-flight check

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/sklearn/manifold/_t_sne.py:783: FutureWarning:

The default initialization in TSNE will change from 'random' to 'pca' in 1.2.


Clustering...
step 0
got 23 at resolution 1.5
step 1
got 15 at resolution 0.75
step 2
got 11 at resolution 0.375
step 3
got 7 at resolution 0.1875
step 4
got 11 at resolution 0.28125
step 5
got 10 at resolution 0.234375
----------

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scmoib-0.0.1+7.gfe906bc.dirty-py3.7.egg/scmoib/tools/metrics/utils/dijkstra.py:137: UserWarning:

Warning! You are running the high load graph method.
Test run information:
CPU: 8 cores
Wall time: 6 minutes
RAM: 4 GB(?)
Number of barcode pairs: 11k

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scmoib-0.0.1+7.gfe906bc.dirty-py3.7.egg/scmoib/tools/metrics/_node_metrics.py:51: RuntimeWarning:

Mean of empty slice.

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/numpy/core/_methods.py:189: RuntimeWarning:

invalid value encountered in double_scalars

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scib-1.0.0-py3.7.egg/scib/metrics/isolated_labels.py:57: DeprecationWarning:

The default dtype for empty Series will be 'object' instead of 'float64' in a future version. Specify a dtype explicitly to silence this warning.

 17%|██████▊                                  | 1/6 [22:51<1:54:18, 1371.60s/it]/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/anndata/compat/__init__.py:183: FutureWarning:

Moving element from .uns['neighbors']['distances'] to .obsp['distances'].

This is where adjacency matrices should go now.


Running pre-flight check
WARNING: .obsp["connectivities"] have not been computed using umap
Clustering...
step 0
got 36 at resolution 1.5
step 1
got 26 at resolution 0.75
step 2
got 18 at resolution 0.375
step 3
got 12 at resolution 0.1875
step 4
got 9 at resolution 0.09375
step 5
got 11 at resolution 0.140625
step 6
got 10 at resolution 0.1171875
----------

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scmoib-0.0.1+7.gfe906bc.dirty-py3.7.egg/scmoib/tools/metrics/utils/dijkstra.py:137: UserWarning:

Warning! You are running the high load graph method.
Test run information:
CPU: 8 cores
Wall time: 6 minutes
RAM: 4 GB(?)
Number of barcode pairs: 11k

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scib-1.0.0-py3.7.egg/scib/metrics/isolated_labels.py:57: DeprecationWarning:

The default dtype for empty Series will be 'object' instead of 'float64' in a future version. Specify a dtype explicitly to silence this warning.

 33%|██████████████▋                             | 2/6 [31:03<56:57, 854.29s/it]
Running pre-flight check

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/sklearn/manifold/_t_sne.py:783: FutureWarning:

The default initialization in TSNE will change from 'random' to 'pca' in 1.2.


Clustering...
step 0
got 21 at resolution 1.5
step 1
got 13 at resolution 0.75
step 2
got 9 at resolution 0.375
step 3
got 12 at resolution 0.5625
step 4
got 11 at resolution 0.46875
step 5
got 10 at resolution 0.421875
----------

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scmoib-0.0.1+7.gfe906bc.dirty-py3.7.egg/scmoib/tools/metrics/utils/dijkstra.py:137: UserWarning:

Warning! You are running the high load graph method.
Test run information:
CPU: 8 cores
Wall time: 6 minutes
RAM: 4 GB(?)
Number of barcode pairs: 11k

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scib-1.0.0-py3.7.egg/scib/metrics/isolated_labels.py:57: DeprecationWarning:

The default dtype for empty Series will be 'object' instead of 'float64' in a future version. Specify a dtype explicitly to silence this warning.

 50%|██████████████████████                      | 3/6 [40:21<35:56, 718.78s/it]/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/h5py/_hl/dataset.py:541: DeprecationWarning:

Passing None into shape arguments as an alias for () is deprecated.


Running pre-flight check

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/sklearn/manifold/_t_sne.py:783: FutureWarning:

The default initialization in TSNE will change from 'random' to 'pca' in 1.2.


Clustering...
step 0
got 21 at resolution 1.5
step 1
got 16 at resolution 0.75
step 2
got 11 at resolution 0.375
step 3
got 8 at resolution 0.1875
step 4
got 9 at resolution 0.28125
step 5
got 11 at resolution 0.328125
step 6
got 9 at resolution 0.3046875
step 7
got 11 at resolution 0.31640625
step 8
got 9 at resolution 0.310546875
step 9
got 10 at resolution 0.3134765625
----------

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scmoib-0.0.1+7.gfe906bc.dirty-py3.7.egg/scmoib/tools/metrics/utils/dijkstra.py:137: UserWarning:

Warning! You are running the high load graph method.
Test run information:
CPU: 8 cores
Wall time: 6 minutes
RAM: 4 GB(?)
Number of barcode pairs: 11k

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scib-1.0.0-py3.7.egg/scib/metrics/isolated_labels.py:57: DeprecationWarning:

The default dtype for empty Series will be 'object' instead of 'float64' in a future version. Specify a dtype explicitly to silence this warning.

 67%|█████████████████████████████▎              | 4/6 [52:16<23:54, 717.30s/it]/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/h5py/_hl/dataset.py:541: DeprecationWarning:

Passing None into shape arguments as an alias for () is deprecated.


Running pre-flight check
Clustering...
step 0
got 18 at resolution 1.5
step 1
got 13 at resolution 0.75
step 2
got 8 at resolution 0.375
step 3
got 11 at resolution 0.5625
step 4
got 10 at resolution 0.46875
----------

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scmoib-0.0.1+7.gfe906bc.dirty-py3.7.egg/scmoib/tools/metrics/utils/dijkstra.py:137: UserWarning:

Warning! You are running the high load graph method.
Test run information:
CPU: 8 cores
Wall time: 6 minutes
RAM: 4 GB(?)
Number of barcode pairs: 11k


KeysView(AxisArrays with keys: X_umap, latent)
Silhouette failed

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scib-1.0.0-py3.7.egg/scib/metrics/isolated_labels.py:57: DeprecationWarning:

The default dtype for empty Series will be 'object' instead of 'float64' in a future version. Specify a dtype explicitly to silence this warning.

 83%|███████████████████████████████████       | 5/6 [1:05:11<12:18, 738.33s/it]
Running pre-flight check

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/sklearn/manifold/_t_sne.py:783: FutureWarning:

The default initialization in TSNE will change from 'random' to 'pca' in 1.2.


Clustering...
step 0
got 21 at resolution 1.5
step 1
got 16 at resolution 0.75
step 2
got 11 at resolution 0.375
step 3
got 8 at resolution 0.1875
step 4
got 9 at resolution 0.28125
step 5
got 10 at resolution 0.328125
----------

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scmoib-0.0.1+7.gfe906bc.dirty-py3.7.egg/scmoib/tools/metrics/utils/dijkstra.py:137: UserWarning:

Warning! You are running the high load graph method.
Test run information:
CPU: 8 cores
Wall time: 6 minutes
RAM: 4 GB(?)
Number of barcode pairs: 11k

/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/scib-1.0.0-py3.7.egg/scib/metrics/isolated_labels.py:57: DeprecationWarning:

The default dtype for empty Series will be 'object' instead of 'float64' in a future version. Specify a dtype explicitly to silence this warning.

100%|██████████████████████████████████████████| 6/6 [1:15:55<00:00, 759.24s/it]
CPU times: user 3h 30min 13s, sys: 1h 20min 46s, total: 4h 51min
Wall time: 1h 15min 55s










    
    
Out[5]:











  
    

      
      conn_ratio
      spec_dist_10
      accuracy
      AMI
      ARI
      NMI
      homogeneity
      graph connectivity broad cell types
      sil_global
      sil_clus
      pairwise_distance
      il_sil
      il_clus
      graph connectivity broad.cell.types
    

  
  
    

      unintegrated
      0.0
      NaN
      0.000000
      0.601217
      0.421081
      0.601592
      0.615105
      0.499798
      0.723077
      0.349730
      0.761411
      NaN
      NaN
      NaN
    

    

      liger
      1.0
      0.629795
      0.322143
      0.532695
      0.448401
      0.533130
      0.551172
      NaN
      0.652315
      0.466547
      0.861671
      NaN
      NaN
      0.633627
    

    

      seurat_cca
      1.0
      0.938760
      0.849163
      0.729525
      0.646358
      0.729793
      0.710242
      0.979320
      0.507945
      0.720455
      0.837083
      NaN
      NaN
      NaN
    

    

      bindsc
      1.0
      0.398853
      0.010280
      0.629792
      0.457215
      0.630131
      0.661708
      0.936445
      0.524120
      0.704856
      0.687280
      NaN
      NaN
      NaN
    

    

      scalex
      1.0
      0.875036
      0.388737
      0.666833
      0.482170
      0.667135
      0.707853
      0.988257
      NaN
      NaN
      0.705320
      NaN
      NaN
      NaN
    

    

      seurat_rpca
      1.0
      0.836273
      0.701055
      0.635951
      0.578401
      0.636312
      0.621872
      0.966974
      0.520412
      0.771678
      0.890710
      NaN
      NaN
      NaN
    

  



























In [6]:



     


scmoib.pl.cumulative_node_group(adata_list[1:], ids[1:])



     






















    
    































In [7]:



     


df = my_obj.get_df()
df



     






















    
    
Out[7]:











  
    

      
      conn_ratio
      spec_dist_10
      accuracy
      AMI
      ARI
      NMI
      homogeneity
      graph connectivity broad cell types
      sil_global
      sil_clus
      pairwise_distance
      il_sil
      il_clus
      graph connectivity broad.cell.types
    

  
  
    

      unintegrated
      0.0
      NaN
      0.000000
      0.601217
      0.421081
      0.601592
      0.615105
      0.499798
      0.723077
      0.349730
      0.761411
      NaN
      NaN
      NaN
    

    

      liger
      1.0
      0.629795
      0.322143
      0.532695
      0.448401
      0.533130
      0.551172
      NaN
      0.652315
      0.466547
      0.861671
      NaN
      NaN
      0.633627
    

    

      seurat_cca
      1.0
      0.938760
      0.849163
      0.729525
      0.646358
      0.729793
      0.710242
      0.979320
      0.507945
      0.720455
      0.837083
      NaN
      NaN
      NaN
    

    

      bindsc
      1.0
      0.398853
      0.010280
      0.629792
      0.457215
      0.630131
      0.661708
      0.936445
      0.524120
      0.704856
      0.687280
      NaN
      NaN
      NaN
    

    

      scalex
      1.0
      0.875036
      0.388737
      0.666833
      0.482170
      0.667135
      0.707853
      0.988257
      NaN
      NaN
      0.705320
      NaN
      NaN
      NaN
    

    

      seurat_rpca
      1.0
      0.836273
      0.701055
      0.635951
      0.578401
      0.636312
      0.621872
      0.966974
      0.520412
      0.771678
      0.890710
      NaN
      NaN
      NaN
    

  



























In [8]:



     


df.drop(columns=['conn_ratio', 'il_clus', 'il_sil'], inplace=True)



     



















In [9]:



     


df



     






















    
    
Out[9]:











  
    

      
      spec_dist_10
      accuracy
      AMI
      ARI
      NMI
      homogeneity
      graph connectivity broad cell types
      sil_global
      sil_clus
      pairwise_distance
      graph connectivity broad.cell.types
    

  
  
    

      unintegrated
      NaN
      0.000000
      0.601217
      0.421081
      0.601592
      0.615105
      0.499798
      0.723077
      0.349730
      0.761411
      NaN
    

    

      liger
      0.629795
      0.322143
      0.532695
      0.448401
      0.533130
      0.551172
      NaN
      0.652315
      0.466547
      0.861671
      0.633627
    

    

      seurat_cca
      0.938760
      0.849163
      0.729525
      0.646358
      0.729793
      0.710242
      0.979320
      0.507945
      0.720455
      0.837083
      NaN
    

    

      bindsc
      0.398853
      0.010280
      0.629792
      0.457215
      0.630131
      0.661708
      0.936445
      0.524120
      0.704856
      0.687280
      NaN
    

    

      scalex
      0.875036
      0.388737
      0.666833
      0.482170
      0.667135
      0.707853
      0.988257
      NaN
      NaN
      0.705320
      NaN
    

    

      seurat_rpca
      0.836273
      0.701055
      0.635951
      0.578401
      0.636312
      0.621872
      0.966974
      0.520412
      0.771678
      0.890710
      NaN
    

  



























In [10]:



     


df.sum(axis=1)



     






















    
    
Out[10]:








unintegrated    4.573011
liger           5.631496
seurat_cca      7.648644
bindsc          5.640680
scalex          5.481340
seurat_rpca     7.159638
dtype: float64
























In [11]:



     


for i,j in zip(adata_list, ids):
    print(j)
    try:
        sc.pl.umap(i, color=['broad cell types', 'omic_batch'], wspace=0.7)
    except:
        sc.pl.umap(i, color=['broad.cell.types', 'omic_batch'], wspace=0.7)



     






















    
    






unintegrated










    
    

















    
    






liger










    
    






/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/anndata/_core/anndata.py:1228: FutureWarning:

The `inplace` parameter in pandas.Categorical.reorder_categories is deprecated and will be removed in a future version. Reordering categories will always return a new Categorical object.

... storing 'orig.ident' as categorical
/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/anndata/_core/anndata.py:1228: FutureWarning:

The `inplace` parameter in pandas.Categorical.reorder_categories is deprecated and will be removed in a future version. Reordering categories will always return a new Categorical object.

... storing 'filter_barcodes' as categorical
/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/anndata/_core/anndata.py:1228: FutureWarning:

The `inplace` parameter in pandas.Categorical.reorder_categories is deprecated and will be removed in a future version. Reordering categories will always return a new Categorical object.

... storing 'louvain2_rna' as categorical
/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/anndata/_core/anndata.py:1228: FutureWarning:

The `inplace` parameter in pandas.Categorical.reorder_categories is deprecated and will be removed in a future version. Reordering categories will always return a new Categorical object.

... storing 'filter' as categorical
/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/anndata/_core/anndata.py:1228: FutureWarning:

The `inplace` parameter in pandas.Categorical.reorder_categories is deprecated and will be removed in a future version. Reordering categories will always return a new Categorical object.

... storing 'omic_batch' as categorical
/home/icb/atai.dobrynin/miniconda3/envs/test2/lib/python3.7/site-packages/anndata/_core/anndata.py:1228: FutureWarning:

The `inplace` parameter in pandas.Categorical.reorder_categories is deprecated and will be removed in a future version. Reordering categories will always return a new Categorical object.

... storing 'group' as categorical










    
    








<Figure size 1088x320 with 0 Axes>










    
    

















    
    






seurat_cca










    
    

















    
    






bindsc










    
    

















    
    






scalex










    
    

















    
    






seurat_rpca










    
    































In [ ]: